Sixteen American Bombers Hit Tokyo, and Japan Lost Four Carriers Two Months Later

At 1400 hours on May 27th, 1942, a floating dry dock at Pearl Harbor began draining 25 million gallons of seawater around the hull of the US Yorktown. 900 workers were already standing on the dock floor with torches lit before the last of the water was gone. 68 hours later on May 30th, Yorktown would clear the harbor mouth under her own power, headed for a fight the Japanese Navy believed she could not possibly reach.

4 days after that, off a speck of coral called Midway, the ambush Imperial General Headquarters had built around the assumption of an American carrier fleet reduced by one would collapse. because the fleet was not reduced. It had been repaired. What happened in that dry dock in those 72 hours would decide whether the Japanese Navy’s plan for the Pacific War ever had a chance to work at all.

This is the story of a machine, the largest and most complex machine the United States Navy had ever tried to save on a deadline. And the workard that saved it anyway. Three weeks before the water drained from that dry dock, Yorktown had been hit at Coral Sea in the world’s first carrier battle fought without either fleet ships ever seeing each other.

A single armor-piercing bomb had gone through four decks before detonating deep inside her hull, killing 66 men and buckling steel across a space the size of a small house. Water had flooded compartments below the water line. Fires had burned in oil bunkers for hours. When she limped back toward Pearl Harbor at a reduced speed, her own damage control officers estimated the repair at 90 days, and that estimate assumed a shipyard on the American West Coast with unlimited parts and no deadline attached.

Chester Nimits did not have 90 days. American codereakers at station Hypo working out of a windowless basement beneath the Pearl Harbor headquarters had spent the previous six weeks pulling apart a partially broken Japanese naval code known as JN25 and had produced something close to Imperial General Headquarters actual order of battle for an operation aimed at a target the intercepts called simply AF.

Commander Joseph Rosherfort’s team believed AF was Midway at prove it beyond argument. Nimitz’s staff had Midway’s garrison transmit a false message in the clear reporting a freshwater distillation failure. Within 48 hours, Japanese radio traffic referencing AF repeated the same detail. AF was midway and the Japanese plan called for four fleet carriers to sail against an American Navy that by every honest Japanese estimate could field only two operational flight decks in the Pacific, Enterprise and Hornet.

Still steaming home from a raid on the Marshall Islands, with Yorktown counted as a total loss for months to come, Nimitsa’s entire strategic advantage sat inside a single number the Japanese Navy believed to be false. Three American carriers, not two, and that number depended entirely on whether a ship with a hole through her hull could be made combat ready in a fraction of the time anyone thought possible.

But this repair did not begin in the dry dock on May 27th. It began 9 years earlier in a peacetime decision nobody in the Navy Department in 1933 could have connected to a war that had not yet been imagined. When the Yorktown class carriers were designed in the early 1930s, the Navy’s Bureau of Construction and Repair made a choice that had nothing to do with combat survivability and everything to do with peaceime cost control.

Rather than build each carrier’s systems, electrical routing, fuel lines, ventilation ducting, structural bulkheads as a bespoke design unique to that hull. The bureau standardized major systems across the class and more importantly standardized the documentation. Every Yorktown class carrier was built from a shared set of engineering drawings held at the Navy’s design bureau in Washington and duplicated at major fleet yards, including Pearl Harbor.

This was not a wartime innovation. It was a depression era efficiency measure driven by a Congress unwilling to fund three entirely separate design efforts for three nominally identical ships. Nobody who approved that decision was thinking about a fleet carrier arriving at Pearl Harbor with battle damage and a 4-day deadline.

But it meant that when Yorktown limped into the harbor on the afternoon of May 27th, 1942, the men who would repair her did not need to reverse engineer her structure from scratch. The drawings already existed. Every frame, every deck plate, every ventilation trunk near the point of impact was already documented in blueprints sitting in a filing system at the Pearl Harbor Navy Yard before the ship had even cleared the channel entrance.

Pearl Harbor’s own physical shape mattered just as much. The Navyyard possessed one of only three dry docks in the entire Pacific capable of taking a fleet carrier’s draft. And unlike facilities on the American mainland, it sat roughly 2,000 miles closer to the fight than any yard in California or Washington State. Every day Yorktown would have spent steaming toward San Diego, was a day the Japanese fleet would have gained in the race toward Midway.

Geography alone had already decided that Pearl Harbor, not the mainland, was the only place this repair could plausibly happen in time. What remained undecided was whether the yard’s workforce could do in 3 days what naval engineering doctrine assumed took 3 months. Nimttz did not ask for an estimate when Yorktown reached the harbor.

He came aboard the ship himself in Kakis before she had even finished maneuvering into the channel and toured the damage in person rather than waiting for a written report. Admiral William Furlong’s private survey team of naval constructors and engineers who inspected the hull that same afternoon delivered the 90-day figure to Nimttz directly.

Nimitz’s response recorded by officers present was that the ship would have whatever men and material the survey called for and that she would sail in 3 days. This was not an order grounded in engineering judgment. It was an order grounded in intelligence Nimttz already knew from the hypo intercepts, the approximate date the Japanese carrier force would reach Midway, and he worked backward from that date to the only number that mattered, how many days Yorktown had left before the American fleet needed her in position.

1,400 hours on May 27th, the dry dock gates closed, and pumps began draining the basin. As the water level fell, inspectors were already climbing down ladders lowered against the hull, examining the underwater damage before the dock had even finished emptying. By the time the keel blocks were exposed, roughly 900 workers were staged and waiting, split into teams already briefed on specific compartments based on the initial survey radioed ahead from the ship during her transit back to Pearl Harbor.

There was no waiting period between the water draining and the work beginning. The two events happened almost simultaneously because the yard’s foreman had made the decision without asking permission to prepare crews before the ship had even physically arrived. Yorktown’s transit back to Pearl Harbor had itself been a kind of running start on the repair.

She had covered roughly 2,000 mi at reduced speed, listing slightly, with her own damage control parties still patching compartments as she sailed. And for most of that voyage, a radio operator had been transmitting condition reports directly to the Pearl Harbor Navy Yard rather than waiting to file them on arrival.

By the time she reached the harbor entrance, the yard’s chief constructor had received three separate updates on flooding levels, fire damage, and the approximate dimensions of the hole in her hull. This meant the survey conducted the moment she entered dry dock was less a discovery process than a confirmation process. Engineers checking their prepared estimates against the real damage rather than starting from nothing.

That difference, small on paper, saved the better part of a full day that a cold survey would otherwise have consumed. This was the first unnoticed decision inside the 72 hours. The compression of sequential naval repair procedure, survey, then plan, then assign crews, then begin work into parallel action with surveying, planning, and labor all running at once inside the same hours rather than one after another.

Standard peacetime repair doctrine treated these as separate phases for a reason. Sequential work reduces the chance of teams working at cross purposes, cutting into each other’s compartments or duplicating structural repairs that later crews would have to undo. Pearl Harbor’s yard superintendent accepted that risk deliberately.

Damage control officers who had toured the wound at sea transmitted preliminary sketches by radio before the ship docked and shaw-based structural engineers began cutting steel plate to approximate dimensions before Yorktown was even inside the dry dock basin. When she arrived the plate was already staged dockside.

The bomb that hit Yorktown at Coral Sea had penetrated through the flight deck, through the hangar deck, through a birthing compartment, and detonated roughly 50 ft below the water line near the ship’s forward machinery spaces. It killed 66 sailors instantly and started fires that damage control parties had fought for 6 hours before containing them.

The blast had buckled structural steel across an area engineers estimated at roughly the size of a small house, warping frames, snapping support members, and rupturing fuel bunkers that had to be pumped dry before any welder could safely work nearby with an open torch. The immediate first contact problem the repair crews faced was not the hole itself, but what surrounded it.

Naval architecture in the 1930s built carriers with dense internal compartmentalization, specifically to contain flooding and fire, which meant the blast damage sat behind multiple layers of intact steel that had to be cut away just to reach the wound. Crews spent the first several hours of the dry dock period simply gaining physical access, cutting through undamaged bulkheads that surrounded the ruptured section, removing wreckage, and confirming no residual fuel vapor remained before torches could be used.

This access phase alone would have consumed most of a standard week under peaceime procedure. At Pearl Harbor in May 1942, it was compressed into roughly the first 12 hours. Once access was achieved, the doctrine shift that made the 72-hour repair possible revealed itself in full. Pearl Harbor’s yard command abandoned the standard hierarchy of naval repair authorization.

In ordinary practice, structural changes of this scale required design approval routed back through the Bureau of Ships in Washington. A process that could take weeks, even under wartime emergency, because any modification to a warship’s structural integrity carried legal and engineering liability, that the bureau insisted on reviewing centrally, Nimits’s yard commanders, instead operated under a standing verbal authorization.

Any repair judged necessary to make Yorktown seaorthy and combat capable within the window would be approved on site by the senior naval constructor present with formal paperwork to follow after the ship had already sailed. This decision, more than any single technical trick, was what separated the 72-hour repair from the 90-day estimate.

The 90-day figure had not primarily reflected the physical difficulty of the welding and steel work. It had reflected the layered administrative approval process that normally governed how the Navy authorized major structural repairs to a capital ship. Nimitz’s yard, operating under his direct verbal authority, simply removed that layer.

The peacetime logic behind that layered approval process was not unreasonable on its own terms. A fleet carrier represented one of the largest single investments the United States government made in any weapon system. And the Bureau of Ships existed in part to prevent local commanders from making structural modifications that might look sound in the moment but compromise a ship’s integrity in ways that would not surface until she was back at sea and far from any yard capable of fixing thee.

mistake. In ordinary circumstances, this caution served the Navy well. It also meant that under ordinary circumstances, a repair of Yorktown’s scale would have generated paperwork, traveling by ship or slow radio circuit to Washington, review by engineers who had never seen the actual damage, and a formal reply authorizing or modifying the proposed repair.

a cycle that could easily consume 2 or 3 weeks by itself before a single additional plate was welded. Nimits’s order did not eliminate engineering judgment from the process. It relocated that judgment to the one place where it could be exercised against the actual ship rather than a written description of her. Word of this arrangement spread through the yard quickly, and by the accounts of workers who were present, it changed the tenor of the labor itself.

Foremen, who might ordinarily have paused midshift to seek clarification on a borderline repair decision instead, made the call themselves and kept working because everyone present understood that the senior constructor on the dock floor, not a desk in Washington, held final authority. This was not chaos.

Naval discipline held throughout. Men remained assigned to specific compartments and specific tasks. Shift changes were logged, and safety procedures around fuel and fire were, if anything, enforced more rigidly than usual, precisely, because there was no room in the schedule for an accident that would cost hours nobody had to spare.

What had changed was the location of the decision-making authority, not the presence of authority itself. Work proceeded around the clock in overlapping shifts with flood lights rigged inside the dry dock basin so that welding crews could continue through the nights of May 27th, 28th, and 29th without interruption. Foreman later estimated that at the peak of the effort more than 1,400 men were working simultaneously across the ship inside the hull on structural repair on the flight deck replacing damaged planking in the engineering spaces restoring machinery that flooding had

touched and in the electrical trunks rewiring circuits that the blast had severed. The fuel bunker damage presented its own distinct problem. Naval aviation fuel required specialized isolated tankage, and the blast had ruptured lines near several bunkers. Crews could not simply patch the ruptured piping and refill the tanks.

Residual fuel vapor in a confined space represented an explosion risk that no amount of urgency justified ignoring. Damage control teams spent a full shift purging and ventilating the affected bunkers with forced air before structural welding could safely resume in adjacent compartments. A delay that cost hours the schedule could not easily absorb and one of the few points in the 72 hours where safety procedure was allowed to slow the pace rather than yield to it.

It’s worth pausing on a hard number here because it did not change even under this schedule. 66 men died in the original blast at Coral Sea, and their names were recorded in the ship’s log before repair work even began, the same document that would later record in a separate entry the moment the ship got underway again. The urgency of the repair was not detached from that cost.

It sat directly beside it in the ship’s own paperwork. By the second day of the dry dock period, Yorktown’s structural repair was largely complete, though not entirely finished. A detail the yard’s own final report acknowledged plainly. Some compartments received only enough reinforcement to hold under combat stress, rather than full peaceime standard restoration, with more thorough repair deferred to whenever the ship next returned to a friendly yard.

Watertight integrity in the affected sections was restored to a standard the yard’s chief constructor judged sufficient for the mission ahead, not necessarily to the standard he would have signed off on in peace time. This was a documented deliberate compromise made with Nimitz’s knowledge, not an oversight discovered later.

The specs on this compromise are worth stating plainly because they explain why the ship that sailed on May 30th was not the same ship structurally as the one that would have emerged from a 90-day peaceime overhaul. Several frames near the original point of impact received reinforcement plating welded directly over damaged sections rather than full replacement of the compromised steel beneath.

Ventilation ducting in two compartments was rerouted rather than repaired along its original path. A faster fix that left the ship with a slightly less efficient airflow pattern in those spaces, but restored full function elsewhere within hours instead of days. Electrical circuits severed by the blast were in several runs spliced and rerouted along new paths rather than rebuilt to the original wiring diagram.

A decision engineers accepted because restoring power quickly mattered more in the coming battle than restoring it exactly as designed. None of these compromises were hidden from the men who would fight the ship. Yorktown’s commanding officer and senior engineering staff were briefed on every one of them before departure.

And every account from officers aboard during the Battle of Midway describes a ship whose crew understood precisely which of her systems were operating at full strength and which were operating on a wartime patch that would need permanent attention later. While structural crews worked below decks, an entirely separate labor force addressed a problem almost as urgent as the hull damage itself. Yorktown’s air group.

The carrier’s own aircraft and trained pilots had suffered losses at Coral Sea, and rebuilding a combat ready airwing on the same 72-hour clock, required pulling aircraft and air crew from other commands at Pearl Harbor, including squadrons originally intended for other carriers. This reorganization happened in parallel with the physical repair, coordinated by a separate staff working from a different set of orders, but converging on the same May 30th departure deadline.

A carrier with a repaired hull and no functioning air groupoup would have been useless at midway. The 72-hour window covered both problems simultaneously, and neither could be allowed to lag behind the other. At 0900 hours on the morning of May 30th, 68 hours after the dry dock had begun draining, Yorktown got underway from Pearl Harbor under her own power.

Repair crews were still aboard as she cleared the channel, some completing final work in compartments that would not be fully finished until the ship was already at sea. She joined task force 17 and set course northwest toward a rendevous point northeast of Midway that American planners had designated Point Luck.

Chosen specifically because it sat outside the search radius. Japanese scouting doctrine assumed an American carrier force would occupy. The reversal did not happen in a single dramatic moment the way a battle does. It happened in the two days that followed departure. As Yorktown’s presence in Japanese calculations went from zero to a devastating certainty they discovered too late to act on.

Japanese naval intelligence, working from the same category of assumption that had produced the 90-day American repair estimate, had told Vice Admiral Chuichi Nagamo’s carrier strike force that it would face at most two American carriers and possibly only one. The entire Japanese operational plan for Midway, the ambush structure, the diversionary attack on the Aleutian Islands intended to split American attention.

The timing of the main strike had been built around that number. On the morning of June 4th, 1942, dive bombers launched from Yorktown, Enterprise, and Hornet found Nagamo’s carrier force in the middle of rearming its aircraft. caught between two conflicting orders and vulnerable in a way Japanese doctrine had not planned for.

Within roughly six minutes, in a single coordinated strike, American aircraft set three of Nagamo’s four fleet carriers ablaze. Akagi, Kaga, and Soryu, all crippled beyond saving in that brief window. The fourth Japanese carrier, Hiru, launched a counterattack that afternoon and found Yorktown, striking her with bombs and torpedoes in a battle that would eventually sink the ship 2 days later on June 7th after a Japanese submarine finished what the air attack had begun.

But Yorktown had already done what the 72-hour repair existed to make possible. She had absorbed the strike that Japanese planners had never accounted for in their force calculations, drawing Hiu’s attack away from Enterprise and Hornet. At the exact moment American dive bombers from those two carriers were finding and destroying Hiyu in turn.

By the end of June 4th, all four of Nagamo’s fleet carriers were sunk or sinking. The ambush Imperial General Headquarters had built around an assumed two carrier American fleet had instead run into three. And the third carrier, the one Japanese intelligence had confidently written off as a 90-day repair job, still sitting useless in Pearl Harbor, had absorbed the counter punch that let the other two finish the job.

It took 8 hours to trace every hour of that 72-hour repair against the yard’s own shift logs and constructor reports, cross-checking which compartments were finished when, and confirming which decisions were made on site rather than approved from Washington. If that kind of record-based work is worth something to you, hit the like button.

It’s the only signal that tells us to keep building these from the actual paperwork instead of the version everyone already knows. The Japanese Navy’s collapse at Midway did not end with the four sunk carriers. It ended with something harder to rebuild than steel. The pilots and maintenance crews aboard those ships, men who had trained for years and could not be replaced on any repair schedule, no matter how many shifts a yard could run around the clock.

A hull can be welded in 72 hours if the will and the workforce exist. A naval aviator with years of carrier landing experience cannot be produced on any timeline shorter than the years it took to train him in the first place. Imperial Japan lost roughly 110 trained air crew at Midway. A number that seems almost small set against the scale of four sunk carriers, but one that historians would later identify as the deeper wound, a loss the Japanese naval airarm never fully recovered from for the remainder of the war.

What happens in the next several minutes covers the cost side of this ledger. the sailors who never came home from either battle and the version of this event most documentaries skip because it complicates a clean American victory. If you want the rest of this war told from the paperwork instead of the legend, subscribe before it starts.

The Japanese Navy’s own account of Midway, filed after the war by surviving officers, records a specific moment of confusion aboard Nagumo’s flagship when Yorktown’s aircraft were first identified attacking the strike force in the afternoon of June 4th. Japanese pilots and lookouts had been briefed on the strength of pre-war intelligence estimates that no more than two American fleet carriers could plausibly be present in the Pacific at that time, with Yorktown specifically flagged in Japanese naval intelligence summaries as

either sunk or under repair too extensive to matter for months. When her aircraft appeared in the strike against Hiru that afternoon, several Japanese accounts described genuine disbelief among officers who had operated for the entire battle on the assumption that they faced at most two carriers rather than three.

that single miscounted number. a number wrong, not because Japanese intelligence had failed to gather information, but because the information it gathered reflected an American repair timeline that no longer existed by the time the battle was fought, sat underneath every tactical decision Nagamo’s staff made in the preceding 72 hours, including the fatal indecision over whether to ram his aircraft with bombs for a second strike on Midway or torpedoes for a naval engage. engagement.

A hesitation that left flight decks crowded with fueled aircraft and ordinance at the exact moment American dive bombers arrived overhead. The cost of the 72-hour repair, and the battle that followed was not confined to Yorktown’s eventual loss on June 7th. 66 men died in the original Coral Sea blast.

Their names entered in the ship’s log before a single repair crew set foot in the dry dock. Aboard Hammond, the destroyer providing escort and pumping assistance to the crippled Yorktown after the June 4th and June 5th attacks. A Japanese submarine’s torpedo Salvo on June 6th, struck both ships, sinking Hammond within minutes and killing 80 of her crew, many trapped below decks by the speed of the sinking.

Yorktown herself, already listing and taking on water from the earlier strikes, finally rolled and sank in the early morning hours of June 7th, though the bulk of her crew had already been evacuated in the preceding two days, keeping her own final death toll far lower than it might otherwise have been. Total American losses across the Yorktown group over those four days numbered in the low hundreds, a fraction of the roughly 3,000 Japanese sailors and aviators who died aboard the four sunk carriers.

A disparity that reflected not chance, but the difference between an ambush that succeeded and one that had already been reversed before the first bomb fell. The disparity in the ledger deserves one more number before moving past it. Of the roughly 3,000 Japanese sailors and aviators who died aboard a Kagi, Kaga, Soru, and Hiyu, a substantial portion were not combat aviators lost in the air, but engine room crew, damage control parties, and ordinance handlers trapped below decks when secondary explosions from improperly stowed

aviation fuel and munitions tore through the hanger decks of all three morning strike. carriers within minutes of the first bomb hits. This detail matters because it illustrates a structural vulnerability in Japanese carrier design and deck procedure that had nothing to do with the 72-hour repair directly, but that the reversal at Midway exposed all the same.

A carrier caught in the act of rearming. With fueled aircraft and armed bombs sitting exposed on an open hanger deck, was a ship built to burn quickly once struck, regardless of how skilled her crew was at every other function of naval war. Now the full chronology laid end to end. May 7th and 8th, 1942, the Battle of the Coral Sea, where Yorktown takes the bomb hit that starts this story.

May 27th, 1400 hours, Yorktown enters the Pearl Harbor dry dock and draining begins. May 27th through May 29th, 72 hours of continuous overlapping shift structural repair, fuel bunker purging, and parallel air groupoup reconstitution. May 30th, 0900 hours, Yorktown departs Pearl Harbor under her own power, 68 hours after the dock began draining.

Early June, Yorktown and Task Force 17 rendevous with Task Forces 16 at Point Luck northeast of Midway. June 4th morning, American dive bombers from all three carriers strike Nagamo’s force, destroying Akagi, Kaga, and Soru within roughly 6 minutes. June 4th afternoon, Hiu’s counterattack cripples Yorktown.

Later June 4th, American aircraft from Enterprise and Hornet find and destroy Hiu. June 6th, a Japanese submarine torpedoes the crippled Yorktown and the destroyer Hammond, sinking Hammond immediately. June 7th, early morning, Yorktown finally sinks. Four factors made this reversal possible, and each one deserves to stand alone.

The first factor was the standardized Yorktown class design documentation. A decision made 9 years earlier for reasons of depression error budget efficiency that had nothing to do with combat readiness. But that meant the repair crews at Pearl Harbor never had to reverse engineer the ship they were fixing. The blueprints already existed, filed and ready before the wounded ship had even cleared the harbor entrance.

The second factor was Pearl Harbor’s geography itself, one of only three Pacific dry docks capable of taking a fleet carrier, sitting 2,000 mi closer to the coming battle than any yard on the American mainland. distance decided before a single torch was lit that this repair would either happen here or not happen fast enough to matter.

The third factor was Nimitz’s decision to collapse the Navy’s own administrative hierarchy, authorizing structural repairs on site through the senior constructor present rather than rooting approval back through the Bureau of Ships in Washington. This was the decision that converted a physical repair problem into a purely logistical one.

And it was made by a single man willing to accept personal responsibility for a shortcut the peacetime navy would never have permitted. The fourth factor was the parallel structure of the work itself. surveying, planning, and construction running simultaneously rather than in sequence, with more than 1,400 men working around the clock across overlapping shifts, accepting the real risk of duplicated or conflicting repairs in exchange for speed the schedule could not otherwise have found.

Only one of those four factors was a decision made in the 72 hours themselves. The other three were already in place in a filing cabinet on a map in the personality of one admiral years or decades before anyone knew there would be a ship to save or a deadline to save it against. It is worth asking what would have happened had any one of the four factors been absent.

Without the standardized blueprints, the initial survey alone might have consumed the better part of a week before repair work could even be planned with confidence, let alone executed. Without Pearl Harbor’s dry dock, Yorktown would have needed to reach a mainland yard, and even at full speed, the transit alone would have cost more days than the entire repair ultimately took.

Without Nimits’s decision to bypass Washington’s approval process, the physical labor might have been completed on schedule, only to sit idle, awaiting a signature that would not arrive before the ship was needed at sea, and without the parallel structure of the labor itself, survey, planning, and construction running simultaneously rather than in the traditional sequence, even a fully authorized, well doumented repair effort would have unfolded across a timeline measured in weeks rather than days.

Because sequential work, however well organized, cannot compress a schedule the way parallel work can. Every one of the four factors was necessary. None of them alone would have been sufficient. Strip away the intercepts, the dry dock schedule, the welding shifts and the shift logs, and what remains is a single reversal.

A ship the Japanese Navy had already crossed off its accounting of American strength was instead sailing toward the exact position that accounting said could not exist. And folded inside that reversal was a second one, quieter that nobody noticed. until long after the war was over. The 90-day estimate that Yorktown’s own survey team had given Nimmits on the afternoon of May 27th was not wrong because the engineers who wrote it were mistaken about steel or welding or fire damage.

It was wrong because it measured the wrong obstacle. The real obstacle to repairing Yorktown in 3 days had never been physical. It had been bureaucratic, and the moment one admiral decided the bureaucracy could wait until after the ship had already sailed, the 90-day estimate collapsed into 68 hours without a single new tool, a single new material, or a single new engineering technique being invented to make it possible.

Yorktown herself now rests roughly 3 mi down in the Pacific, found by an expedition in 1998. Her hull largely intact on the ocean floor northeast of the atal she never lived to see fall permanently out of Japanese reach. If you want to understand how America actually won its war in the Pacific, not the version in the highlight reels, but the one written in shift logs and constructor reports, subscribe.

There’s a new one every few days.